The project complements the human studies of osteoporosis pathophysiology by addressing fractures, the principle clinical manifestation of osteoporosis. Through the unique data resources of the Rochester Epidemiology Project, we can identify large inception cohorts of Olmsted County (or Rochester), Minnesota, residents with specific medical and surgical conditions and conduct a series of retrospective (historical) cohort studies to estimate the long-term risk of age-related fractures associated with secondary osteoporosis, an important contributor to bone loss in the elderly. Secondary osteoporosis is an important area of research because there are therapies for affected men and older women who are not candidates for estrogen treatment. We previously determined the risk of fracture among cohorts of patients with hyperparathyroidism, thyroidectomy, asthma, gastrectomy, inflammatory bowel disease, pernicious anemia, infertility, oophorectomy, orchiectomy, end-stage renal disease, urolithiasis, stroke, venous thromboembolism and anticoagulant therapy, anorexia nervosa, dementia, parkinsonism, epilepsy, poliomyelitis, rheumatoid arthritis and ankylosing spondylitis. We now propose to extend this work by quantifying the fracture risk associated with conditions that might reflect altered sex steroid levels in older women and men including putative elevation (prostate and breast cancer, adult-onset diabetes mellitus in women), suppression (specific disease treatments, diabetes mellitus in men) or both (various indications for hysterectomy). Thus, each condition represents a natural experiment with respect to the pathogenesis of osteoporosis. These represent the only assessments of overall fracture risk among large cohorts of unselected patients from the community, and the results should be more valid and more precise than previous estimates. Our overall goal is to develop new information that will lead to effective strategies for preventing osteoporosis-related fractures among the elderly.This project contributes by demonstrating the public health significance of specific conditions with respect to fractures, by testing important hypotheses and by identifying high-risk subgroups within each cohort. This information is urgently needed by patients and their physicians and also may provide a basis for generating hypotheses to be tested in future research.